DeepSeek Raises V4 Model Prices Ahead of Potential IPO

DeepSeek will significantly increase prices for its flagship V4 AI models starting August 16, marking a strategic shift away from its role as a low-cost provider. The Hangzhou-based company is implementing a dynamic pricing structure designed to manage resource allocation and incentivize users to move workloads to off-peak hours.

Under the new system, peak-hour rates for the DeepSeek-V4-Flash model will rise from $0.28 to $1.32 per 1 million output tokens. The V4-Pro model will see a steeper increase, moving from $0.87 to $3.96 per 1 million tokens during peak periods. Rates during off-peak hours will be set at half of the peak price.

DeepSeek is making these changes as it undergoes massive fundraising and prepares for a potential initial public offering later this year. Founder Liang Wenfeng is balancing the high capital demands of computing infrastructure with the expectations of new investors.